cosmosConnector.py

# cosmosConnector.py
from jsonschema import ValidationError
from langchain_openai import AzureOpenAIEmbeddings
from models.converterModels import PowerConverter
from models.converterVectorStoreModels import PowerConverterVector
import os
from azure.cosmos import CosmosClient
from typing import List, Optional, Dict
from rapidfuzz import fuzz
import logging
import os
from dotenv import load_dotenv
from semantic_kernel.functions import kernel_function

load_dotenv()
# Initialize logging
logger = logging.getLogger(__name__)

class CosmosLampHandler:
    
    def __init__(self, logger: Optional[logging.Logger] = None):
        self.client = CosmosClient(
            os.getenv("AZURE_COSMOS_DB_ENDPOINT"),
            os.getenv("AZURE_COSMOS_DB_KEY")
        )
        self.database = self.client.get_database_client("TAL_DB")
        self.container = self.database.get_container_client("Converters_with_embeddings")
        self.logger = logger
        self.embedding_model = AzureOpenAIEmbeddings(
            azure_endpoint=os.environ["OPENAI_API_ENDPOINT"],
            azure_deployment=os.environ["OPENAI_EMBEDDINGS_MODEL_DEPLOYMENT"],
            api_key=os.environ["AZURE_OPENAI_KEY"]
        )

    async def _generate_embedding(self, query: str) -> List[float]:
        """Generate embedding for the given query using Azure OpenAI"""
        try:
            return self.embedding_model.embed_query(query)
        except Exception as e:
            logger.error(f"Embedding generation failed: {str(e)}")
            raise

    async def get_compatible_lamps(self, artnr: int) -> List[str]:
        """Get compatible lamps for a converter with fuzzy matching"""
        try:
            parameters = [{"name": "@artnr", "value": artnr}]
            query = "SELECT * FROM c WHERE c.artnr = @artnr"
            
            # Collect results properly
            results = [item for item in list(self.container.query_items(
                query=query,
                parameters=parameters
            ))]
            
            if not results:
                return []
                
            return list(results[0]["lamps"].keys())
        
        except Exception as e:
            logger.error(f"Failed to get compatible lamps: {str(e)}")
            return []
        
    async def get_converters_by_lamp_type(self, lamp_type: str, threshold: int = 75) -> List[PowerConverter]:
        """Get converters with fuzzy-matched lamp types"""
        try:
            # Case-insensitive search with fuzzy matching
            query = """
            SELECT
                *
            FROM c WHERE IS_DEFINED(c.lamps)"""
            converters = []
            results = list(self.container.query_items(
                                                    query=query,
                                                    enable_cross_partition_query=True))
            for item in results:
                lamp_keys = item.get("lamps", {}).keys()
                matches = [key for key in lamp_keys 
                          if fuzz.ratio(key.lower(), lamp_type.lower()) >= threshold]
                
                if matches:
                    converters.append(PowerConverter(**item))
            
            return converters
            
        except Exception as e:
            logger.error(f"Lamp type search failed: {str(e)}")
            return []

    
    async def get_lamp_limits(self, artnr: int, lamp_type: str) -> Dict[str, int]:
        """Get lamp limits with typo tolerance"""
        try:
            parameters = [{"name": "@artnr", "value": artnr}]
            query = """
            SELECT c.lamps FROM c 
            WHERE c.artnr = @artnr
            """
            results_iter = list(self.container.query_items(
                query=query,
                parameters=parameters
            ))

            results = [item for item in results_iter]  # Collect results asynchronously

            if not results:
                return {}

            lamps = results[0]["lamps"]

            # Fuzzy match lamp type
            best_match = max(
                lamps.keys(),
                key=lambda x: fuzz.ratio(x.lower(), lamp_type.lower())
            )

            if fuzz.ratio(best_match.lower(), lamp_type.lower()) < 65:
                raise ValueError("No matching lamp type found")

            return {
                "min": int(lamps[best_match]["min"]),
                "max": int(lamps[best_match]["max"])
            }

        except Exception as e:
            logger.error(f"Failed to get lamp limits: {str(e)}")
            raise
    


    async def query_converters(self, query: str) -> str:
        try:
            print(f"Executing query: {query}")
            items = list(self.container.query_items(
                query=query,
                enable_cross_partition_query=True
            ))
            print(f"Query returned {len(items)} items")
            items = items[:10] 
            # self.logger.debug(f"Raw items: {items}")

            items = [PowerConverter(**item) for item in items] if items else []

            self.logger.info(f"Query returned {len(items)} items after conversion")            

            return str(items)
        
        except Exception as e:
            self.logger.info(f"Query failed: {str(e)}")
            return f"Query failed: {str(e)}"
    
    
    async def RAG_search(self, query: str, artnr: Optional[int] = None, threshold: int = 75) -> List[PowerConverterVector]:
        """Hybrid search using raw Cosmos DB vector search"""
        try:
            # Generate embedding
            print(f"Performing hybrid search for query: {query} (ARTNR: {artnr})")
            query_vector = await self._generate_embedding(query)
            
            sql_query = """
                SELECT TOP 5
                    c.id,
                    c.converter_description,
                    c.ip,
                    c.efficiency_full_load,
                    c.name,
                    c.artnr,
                    c.type,
                    c.lamps,
                    c.pdf_link,
                    c.nom_input_voltage_v,
                    c.output_voltage_v,
                    c.unit,
                    c["listprice"],
                    c["lifecycle"],
                    c.size,
                    c.ccr_amplitude,
                    c.dimmability,
                    c.dimlist_type,
                    c.strain_relief,
                    c.gross_weight,
                    VectorDistance(c.embedding, @vector) AS SimilarityScore
                FROM c 
                ORDER BY VectorDistance(c.embedding, @vector)
                """
            
            parameters = [{"name": "@vector", "value": query_vector}]

            # Execute query
            results = list(self.container.query_items(
                query=sql_query,
                parameters=parameters,
                enable_cross_partition_query=True
            ))
            
            items = []
            for item in results:
                items.append(item)
         
            converters = []
            for item in items:
                # Convert float values to integers before validation
                if "lamps" in item:
                    for lamp_key in item["lamps"]:
                        lamp_data = item["lamps"][lamp_key]
                        lamp_data["min"] = int(lamp_data["min"])
                        lamp_data["max"] = int(lamp_data["max"])
  
                converters.append(PowerConverterVector(**item))
                   
            return converters
        except ValidationError as exc:
            print(exc)
                
        except Exception as e:
            logger.error(f"Hybrid search failed: {str(e)}")
            print(f"Hybrid search failed: {str(e)}")